Molecular Dynamics for Prediction of RTs

Accurate prediction of chromatographic retention remains a central challenge in liquid chromatography (LC), despite rapid advances in artificial intelligence and machine learning. This presentation begins by revisiting the fundamental physico‑chemical origins of LC retention, exposing why purely data‑driven AI/ML models based on molecular descriptors and mere experimental conditions often fail to capture the true complexity of analyte–stationary phase interactions in reversed-phase LC and in anion-exchange - RP mixed mode chromatography. By introducing molecular dynamics simulations (MDS), we explore a physics‑based framework capable of resolving the liquid‑to‑solid adsorption phenomena that govern retention behavior of both neutral and charged analytes. Realistic molecular‑level insights reveal how surface solvation, surface heterogeneity, and transient interactions shape chromatographic outcomes beyond what mere structural descriptors can describe. The talk demonstrates how integrating MDS‑derived features with AI/ML can dramatically enhance prediction accuracy and robustness. This hybrid paradigm offers a transformative pathway toward more reliable retention modeling, method development, and smarter chromatographic design in pharmaceutical analysis.